UKNNL holds around 100 million files dating back to the 1940s, and a new artificial intelligence (AI) enabled search platform now finds what teams need at least 10× faster than before, surfacing past research that is directly shaping cutting-edge programmes of the future.
Challenge
The UK’s nuclear sector holds one of the world’s largest and most complex technical archives. UKNNL alone stores around 100 million electronic files, some originating from as long ago as the 1940s.
Conventional search methods make it difficult and time-consuming to locate specific information, especially as many original researchers are no longer available to provide context.
This creates risks of duplicated work, programme delays and loss of valuable expertise. As the UK develops next-generation reactors and fuel-cycle technologies, rapid access to historical knowledge is critical.
Solution
To support the streamline searching of historical nuclear knowledge, a UKNNL team identified, tested and proved an AI-enabled natural language search platform and deployed this on its secure networks to index and interpret millions of scientific documents.
Deployment of the tool at this scale, with the support of the Office for Nuclear Regulation, represents the first of its kind both in the nuclear sector and internationally. It enables:
- search speeds at least 10× faster than manual methods
- more comprehensive and relevant search results
- discovery of previously overlooked or inaccessible content
- improved cross‑sector collaboration
- enhanced information security through identification of legacy data AI-enabled searching is now actively supporting research and development, programme design and technical decision-making across UKNNL and partner organisations.
AI-enabled searching is now actively supporting research and development, programme design and technical decision-making across UKNNL and partner organisations.
Using knowledge to support growth: Developing the UK civil nuclear library
As part of UKNNL’s commitment to unlocking the full value of its knowledge assets, the organisation has established a comprehensive Knowledge Asset Management Strategy. This strategy sets out how UKNNL will identify, protect and maximise the impact of its extensive historical and contemporary knowledge base.
Working closely with industryleading experts, and supported by advanced AI‑enabled search tools, UKNNL is developing structured libraries of information that can be shared with partners to accelerate the deployment of advanced nuclear technologies.
We have already assembled a curated catalogue of more than 50 topic areas, spanning the entire nuclear fuel cycle. This provides a clear framework for communicating, accessing and exploiting knowledge assets across the sector and forms an essential building block for the emerging UK civil nuclear library.
“ AI-enabled searching has transformed our ability to understand and exploit our wealth of knowledge assets. ”
Luke O’Brien
Fellow, UKNNL
Thermodynamic modelling support: Streamlining research into next-generation fuels
Crucial work is under way at UKNNL’s Preston Laboratory to convert highassay low-enriched uranium (HALEU) from its gaseous form into chemically stable compounds that can be used to make fuel for the next generation of advanced nuclear reactors.
Using the AI search engine, the team uncovered a kinetic study conducted in 1962 that was not publicly available elsewhere.
This document confirmed the predicted product and provided insights which directly shaped the design and statistical analysis of upcoming experiments, strengthening confidence in the approach. Without using the AI search engine, reproducing the equivalent data in the paper would have required months, or even years, of additional work to progress the project.
The use of AI on the HALEU project shows just how these tools can support the efficiency and effectiveness of the teams working to build a nuclear future.
“ The discovery of this paper through the use of AI searches has been extremely useful to validate and support the use of thermodynamic modelling for this project. It gave our team a real confidence boost in our results and methods to look at novel systems moving forward. ”
Sophie Cooper
Chemical Modeller, UKNNL
Digital technology in action: Capturing knowledge about advanced modular reactors
To support cross-sector collaboration in the development of advanced modular reactors (AMR), it is essential to be able to leverage the UK’s AMR knowledge base.
Funded by the UK Department for Energy Security and Net Zero (DESNZ) and delivered by UKNNL alongside Arup, AMR Knowledge Capture Programme digitised and indexed this knowledge to improve its accessibility.
More than 7,000 paper documents were scanned, indexed and integrated into a searchable SharePoint platform using AI-generated metadata and natural language processing.
This work prevents the loss of critical historical knowledge and supports faster, more cost‑effective AMR development.
DESNZ confirmed that the project laid a strong foundation for improving accessibility of historical nuclear information, helping to position the UK as a leader in next-generation nuclear technology development.
What the future holds
UKNNL is continuing to expand and refine its knowledge assets, working towards the development of a comprehensive UK civil nuclear library.
AI-enabled search will play a central and increasing role in helping to improve technical decision-making and reduce duplication of effort.
It will also be invaluable in supporting national programmes and training the future nuclear workforce, all of which can only serve to strengthen the UK’s position in advanced nuclear technologies.
“ AI-driven search is transforming how nuclear knowledge is accessed and used. Through the UK’s AMR Knowledge Capture project, we have shown how complex legacy information can be analysed and made accessible beyond conventional methods, unlocking new insights, reducing rework and accelerating cost-effective innovation across the sector. It also enhances collaboration by enabling knowledge to be more readily shared and applied. ”
